Responsible for the MLOps course of the Masters of Science in Data Science. Created and taught the course material and homework projects.

Responsible for the Deep Learning course of the Big Analytics and Metrics Expert master. Created and taught the course material and homework projects.

- Responsible for the ML roadmap and cross-team dependencies.- Defined and organized the hiring process for the Machine Learning team.- Organized and drove the fair Machine Learning effort.

- Text classification: Developed a job roles identification system using textual data from candidates' profiles and resumes, with accuracies above 95 %. Leveraged modern Natural Language Processing techniques (Transformers, BERT) and handled the deployment (FastAPI, SageMaker) and monitoring (Rollbar, New Relic) of the corresponding model.- Fair Machine Learning: designed fairness metrics dashboards to monitor bias on the platform, and added bias mitigation procedures to our model training scripts.- Ranking: rebuilt a ranking algorithm deployment process with GitHub Actions, allowing for gradual roll outs and A/B tests.

- Improved and maintained the applicants' evaluation algorithms (Tensorflow, SageMaker). Developed a real-time system with latencies under 200 ms as a replacement for the existing batch scoring system. Continuously developed new features and models with increased accuracy for every model generation. - Lead the creation of a monitoring and alerting system for the applicants' evaluation system, allowing us to catch errors early and have zero severe failures in 2020. - Co-invented a fast method for actionable feedback against machine learning models' decisions, leveraging Generative Adversarial Networks. Published and presented two research papers and filed a patent.

Laid the business analytics and machine learning foundations of Visage, a recruitment platform powered by crowdsourcing and AI. - Job titles similarity: designed, implemented and deployed a Siamese Neural Network that recommends the most relevant candidates for a given job. Automates 20 % of the candidates' evaluations (TensorFlow, Gensim, Docker, Jenkins).- Resume/job description matching: wrote a machine learning solution with automated training to evaluate candidates for a given job. This algorithm assesses up to 50 % of the submitted candidates' profiles (Scikit-learn, Spacy, AWS ECR).- Business Analytics: determined the relevant KPIs and built the corresponding dashboards (Tableau Software, MongoDB)

Instructor of the 8-weeks Data Analytics for Manager course. The curriculum includes classes on web analytics, A/B testing, SQL, statistics, data visualization, machine learning, and big data.I wrote the class notes, created interactive exercises, and revamped the curriculum and corresponding slides.

Data analysis and predictive models implementation on small and big data architectures. Dozen of missions in various sectors: insurance, healthcare, telecommunications, web, music industry, public sector.Most relevant experiences:– Built a customers' emails dispatch program using topic mining and sentiment analysis, leading to better prioritization and improved customer satisfaction.– Led the design and implementation of a chatbot that automates 80 % of the first messages of a customer service in the education space, thus allowing the agents to focus on the more complexes conversations.– Optimized the prioritization strategy of an insurance call center with machine learning algorithms, leading to a revenue increase of 150k euros per year. – Created a MOOC on Big Data (EIVP) and gave lectures on Deep Learning (UTT).